- Email: [email protected]
Mechanistic study for the formation of polyoxymethylene dimethyl ethers promoted by sulfonic acid-functionalized ionic liquids
Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical journal homepage: www.elsevier.com/locate/molcata
Mechanistic study for the formation of polyoxymethylene dimethyl ethers promoted by sulfonic acid-functionalized ionic liquids Fang Wang, Gangli Zhu, Zhen Li, Feng Zhao, Chungu Xia ∗ , Jing Chen State Key Laboratory for Oxo Synthesis and Selective Oxidation, Suzhou Research Institute of LICP, Lanzhou Institute of Chemical Physics (LICP), Chinese Academy of Sciences, Lanzhou 730000, PR China
a r t i c l e
i n f o
Article history: Received 5 February 2015 Received in revised form 14 July 2015 Accepted 24 July 2015 Available online 31 July 2015 Keywords: Polyoxymethylene dimethyl ethers Sulfonic acid-functionalized ionic liquids Mechanism DFT
a b s t r a c t Polyoxymethylene dimethyl ethers (DMMn ), which are ideal additives for diesel fuel, are mainly synthesized from the condensation of methanol (MeOH) or dimethoxymethane (DMM) with 1,3,5-trioxane (TOX) or paraformaldehyde (PF) promoted by different acid catalysts. However, up to date, few studies have been reported to examine the formation mechanism of DMMn which is essential in understanding the reaction and valuable in designing improved catalysts. In this work, using the density functional theory (DFT) calculations combined with experiment studies, we evaluate the formation mechanism of DMMn which is promoted by sulfonic acid-functionalized ionic liquids (SO3 H-FILs). Our calculated results indicate TOX and PF should dissociate into formaldehyde monomers firstly and then to react with MeOH or DMM. However, their decomposition process is different where the dissociation of TOX proceeds along a two-step mechanism while it follows a one-step mechanism for PF dissociation. As for the formation of DMMn , the reaction proceeds along a hemiacetal-carbocation pathway when MeOH is selected as the capping group provider, while the reaction follows a carbocation pathway when DMM is chosen. The origination for the product distribution pattern has also been discussed in detail. The cations and anions of ionic liquids are found synergistically promote the condensation reaction by proton transfer and simultaneously stabilizing the formed intermediated and transition states. Moreover, all the processes related to the decomposition of TOX and PF and the condensation reaction are reversible. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Polyoxymethylene dimethyl ethers (DMMn , generally 1 ≤ n ≤ 8) are a series chain compounds with the formula CH3 O(CH2 O)n CH3 containing high oxygen contents (42–51%) and cetane numbers (CN). Especially, the CN of DMMn with n = 3–8 is up to 76 or more and the oxygen contents is between 47% and 50%, making them be considered as the promising green additives to diesel fuel, which can lead to the improved thermal efficiency, and the significantly reduced emissions of particular materials, such as COx and NOx . Therefore, the synthesis and application of DMMn have attracted worldwide attentions [1–13]. Two type raw materials are required for the synthesis of DMMn : one refers to a compound providing CH2 O group as a chain segment, such as 1,3,5-trioxane (TOX), paraformaldehyde (PF), or formaldehyde (FA), and the other is a compound which offers CH3 -capping group, such as methanol (MeOH), dimethoxymethane (DMM), or dimethyl ether (DME).
∗ Corresponding author. Fax: +86 931 4968129. E-mail address: [email protected] (C. Xia). http://dx.doi.org/10.1016/j.molcata.2015.07.028 1381-1169/© 2015 Elsevier B.V. All rights reserved.
Both these two type compounds can be synthesized from MeOH which mainly produced from coal and natural gas and facing seriously overplus. Thus, the synthesis of DMMn also promotes the transformation of MeOH into clean energy. So far, many catalysts have been developed for the condensation reaction, such as liquid acids [1], solid superacid [2,3], ion exchange resin [4–10], molecular sieves [11,12], metal oxide [13] and so on. However, few catalysts are suitable for industrial application, so the exploration of effective and environmentally benign catalysts for the acetalization process needs to be solved urgently as oil resources are dwindling gradually. In recent years, our group have developed several sulfonic acid-functionalized ionic liquids (SO3 H-FILs) which are found very effective in promoting the condensation reaction [14,15]. The consecutive research found this catalyst possess many advantages, such as enhance the conversion of raw materials up to 90%, increase the selectivity for DMM3–8 up to 40% and so on, which suggest the condensation reaction catalyzed by SO3 H-FILs exhibiting promising application prospect. For the formation of DMMn , there are several key points we concerned: (1) do TOX and PF decompose firstly before they reacting with MeOH or DMMn ? (2) how does the chain length actually
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
propagate? (3) do different capping group providers, such as MeOH and DMM, lead to distinct reaction pathways? (4) what roles do the catalysts indeed play in the reaction? However, to the best of our knowledge, the published articles and patents are concentrated on the synthesis of DMMn , the development of catalysts [3,4,10], reaction kinetics [9,13,16,17], and chemical process design [18–22]. Few computational studies have been conducted to investigate the questions listed above to date at a molecular level for this important reaction regardless of which kind catalysts being used. In particular, employing experimental methods Zhao et al. [23] have found the growth of DMMn chain is carried out via addition FA one by one and the products distribution follow the Schulz–Flory law which restricts the increase of the selectivity to the desired products. However, the detailed reaction pathways are still unknown. Zheng et al. based on their experiment studies proposed DMMn + 1 are probably formed from the directly addition of protonated FA into C O bond of DMMn [10]. However, the proposed mechanism has not been validated no matter by computational analysis or experimental methods. As for the possible form of TOX or PF to react with end-capping species, Burger et al. have proposed TOX may decompose into a linear trioxymethylene chain firstly and then this chain complex directly inserted into another DMMn leading to the formation of DMMn+3 , such as DMM4 , DMM7 and so on. But they didn’t observe the favorite formation of DMM4 and DMM7 in experiments [9,24]. To clarify the problems listed above, we have carried out a detailed density functional theory (DFT) study for the reactions of TOX/PF with MeOH and with DMM, as depicted in Scheme 1. Based on previous reports of our group [14], 1-(4-sulfonic acid)-butyl-3methylimidazolium hydrogen sulfate [BsMim][HSO4 ] are chosen as the catalysts. We expect the present results will provide new insights into the formation mechanism of DMMn and explain the previous experimental observations in some extent.
229
equipped with a flame ion-ization detector (FID) and a SE-54 capillary column (60 m × 0.25 mm × 0.25 m). High purity nitrogen was used as the carrier gas. It is worth noting that, to avoid further process of the reaction, the ionic liquids have been separated from the sample before GC analysis. Firstly, acetic ether (2.0 g), internal standard (furanidine, 0.02 g), and small amounts CaO (which was used to neutralize any possible acids in the sample mixture) were put into a vial. Then the reaction sample (about 0.2 g) was rapidly transferred to the vial. Thus, ionic liquids were separated owing to its insolubility in acetic ether. The unconverted FA is ascertained by titration method. Two drops of thymol phthalein indicator were added into the sodium sulfite solution (1 mol/L, 20.0 g) firstly, and then this solution was neutralized to colorless by sulphuric acid. Next, the reaction sample (1.0 g) was added into the obtained solution and shocked to blue. At the end, the content of FA can be determined by the titration of sulphuric acid (0.1 mol/L). Molar distribution of every component was adopted to clearly show the content of DMMn yielded from the decomposition of DMM3 , which is defined as md (n, t) =
mDMMn /MDMMn mtotal
where mtotal = the sum mass of all components which is set as 100 g, mDMMn = the mass of DMMn , MDMMn = molar mass of DMMn . The unit of molar distribution is mol/100 g. During the decomposition of DMM3 , the total mass of all components is a constant while the total mole amounts of DMMn keep varying. To clearly reflect the changes of the individual component, the total mass of DMMn was chosen as the denominator. It is worth noting that the molar distribution (mol/g) is not given as the complement to 100% while the total mass fraction (g/g) would be 100%. 3. Results and discussion
2. Computational and experimental details 3.1. The decomposition of TOX and PF 2.1. Computational methods All the DFT calculations were carried out with the Gaussian 09 software package [25]. The hybrid GGA functional B97D [26] which includes dispersion corrections combined with 6-31+G(d,p) [27–30] basis set was selected for the geometry optimization. All geometries were fully optimized without any symmetry constraints. The geometries for the transition states were located employing the synchronous transit-guided quasi-Newton (STQN) method [31] and Berny algorithm [32]. Vibrational frequency analysis was performed to confirm the nature (minima or first-order saddle points) of the stationary points at the same level. Zero-point energy corrections were carried out for all calculated energies. The intrinsic reaction coordinate (IRC) [33] was conducted in both directions (forward and reverse) from the transition states to the corresponding local minima to identify the minimum-energy paths. To take the entropy effects into account, the Gibbs free energies (G) were used in the following discussion. 2.2. Experimental details [BsMim][HSO4 ] were prepared according to the similar procedure reported previously [34,35]. DMM3 (99.9%, technical) was separated from DMMn mixtures. For the decomposition of DMM3 , DMM3 and [BsMim][HSO4 ] catalyst with a mass ratio of 50:1 were charged into a 100 mL glass reactor. The reaction was performed with rapid agitation by magnetic stirrer at 30 ◦ C and 0.1 MPa. The products were identified and quantitatively analyzed by gas chromatography (GC, Agilent 6890)
Before analyzing the reaction mechanism, the chain segment provider TOX and PF dissociate or not with the assistance of SO3 HFILs were firstly examined. This will be helpful to understand how DMMn chains increase in the following steps. 3.1.1. TOX decomposition As can be seen in Fig. 1, beginning with the SO3 H-FIL + TOX, a two-body complex IM1 is firstly formed via hydrogen-bonding interaction between H atom of SO3 H moiety and one O atom of TOX. Subsequently, IM1 can go through TS1–2 to open the six-membered ring and produce a linear trioxymethylene chain complex where the carbocation side is stabilized by the O atom of HSO4 − anion (see IM2). This process is carried out via the migration of H atom of SO3 H to O atom of TOX making one C O bond of TOX be broken. 20.22 kcal mol−1 is required to conquer the energy barrier. It is worth to note that the geometry of trioxymethylene chain is very similar to that of hemiacetal. And then the transferred hydrogen atom goes back to SO3 H-FIL making the decomposition of TOX, i.e., the formation of monomeric FA, as the large C O bond distances 2.072 and 1.968 Å have been illustrated in IM4. The saddle point about this step is TS3–4 with a barrier of 13.45 kcal mol−1 . The separation of monomeric FA from SO3 H-FIL, which is exothermic by 15.18 kcal mol−1 , completes the whole process. According to the results discussed above, it is obviously found that the decomposition of TOX into FA monomers occurs via two steps: (i) the protonation of TOX by SO3 H-FILs resulting in the six-membered ring opening and the formation of a linear trioxymethylene chain complex, (ii) the decomposition of the chain complex to produce
230
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
Scheme 1. The reaction of TOX/PF with CH3 OH and DMM.
the monomeric FA. The rate-determining step refers to the first step, i.e., the protonation of TOX. Additionally, taking free SO3 H-FILs and monomeric FA as the starting point, the reverse reaction can also occur. It overcomes a relatively higher barrier 22.24 kcal mol−1 via TS3–4 forming a trioxymethylene chain complex and then surmounts a little lower energy barrier 20.16 kcal mol−1 via TS1–2 leading to the regeneration of TOX. Apparently, the first step which is more energy demanding is the rate-determining step for the reverse reaction. A little higher energy barrier by 2.02 kcal mol−1 compared with the forward reaction suggests although the decomposition of TOX is reversible, the forward reaction, i.e., the formation of monomeric FA, is more favorable. Another point we interested is if the linear trioxymethylene chain can further react with MeOH or DMM directly leading to the formation of DMM4 . To clarify this question, the corresponding mechanistic details have also been investigated. We first explored
the capping process of linear trioxymethylene by MeOH (see Fig. S1 in Supporting information). The calculated results indicate the energy should be surmounted is as high as 49.47 kcal mol−1 , indicating this process is difficult to realize in mild heating conditions. For the reaction of trioxymethylene chain with DMM, the calculated energy barrier is 51.53 kcal mol−1 , demonstrating this process is also not feasible. Both these results suggest the formation of DMM4 via trioxymethylene chain reacting with MeOH or DMMn directly will be negligible, which is also not observed in experiment [24]. Based on the results discussed above, it is noticeably found that SO3 H-FILs plays important roles in the decomposition process of TOX. On the one hand, hydrogen transfer along hydrogen bonds between cation and TOX can promote the ring opening and the decomposition of TOX. On the other hand, the formed carbocation trioxymethylene chain can be stabilized by O atoms of HSO4 − anion. In other words, the cation and anion can synergistically promote the TOX decomposition.
Fig. 1. Energy surface for the decomposition of 1,3,5-trioxane (TOX) catalyzed by SO3 H-FILs and corresponding optimized structures of intermediates and transition states. The summary energy of free TOX and SO3 H-FILs is taken as the zero point energy. Distances are in Å and energies are in kcal mol−1 . For clarity, the structures of imidazolium and alkyl side chain are depicted in line.
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
Scheme 2. PF decomposition.
3.1.2. PF decomposition After clarifying the decomposition mechanism of TOX, our attentions turns to the dissociation of PF. Considering the calculation amount, the degree of PF polymerization we chosen is 3. Being different from the structure of TOX, the ends of PF contain an H atom and an OH group, respectively, which make the decomposition of PF liberate a H2 O molecule, as depicted in Scheme 2. At the reaction entrance, PF binding to SO3 H-FILs via strong hydrogenbonding interactions results in the formation of IM5 (see Fig. 2), which is energetically below the entrance by 5.96 kcal mol−1 . Subsequently, H1 atom of SO3 H group transfers to the oxygen atom of one OH group resulting in the liberation of H2 O, as depicted by TS5–6 in Fig. 2. Simultaneously, H2 atom from the other OH group of PF transfers to HSO4 − anion, and H3 atom of HSO4 − anion shifts to one oxygen atom of SO3 group. All of these processes contribute to the one-step decomposition of PF and the regeneration of SO3 HFILs catalyst. The energy barrier of this step is 23.89 kcal mol−1 , which is easily to be overcome under mild heating condition. The moderate hydrogen-bonding interaction between monomeric FA and SO3 H-FILs make the catalyst dissociation very facile, which is exothermic by 12.52 kcal mol−1 . The barrier for the backward reaction is 28.16 kcal mol−1 , which is higher than the forward process by 4.27 kcal mol−1 , implying the regeneration of PF is energetically less favorable. However, the barriers 23.89 and 28.16 kcal mol−1 is not too high to be overcome under heating conditions. Thus, the decomposition of PF is also an equilibrium reaction. Additionally, from the geometries of TS5–6 , it is obviously found that SO3 H-FILs play indispensable roles for PF decomposition via protons transfer. The liberation of SO3 H-FILs accomplished the whole process. From the discussion given above, it is clear that the dissociation of TOX proceeds along a two-step pathway, while it is realized via one step for PF decomposition. Moreover, both the decomposition of TOX and PF is facile to proceed under acid existing and mild heating conditions, suggesting they may firstly dissociate into FA monomers and then to react with MeOH or DMM. Thus, FA instead of TOX and PF is used directly in the following discussions about the mechanistic details of the DMMn formation process. In addition, it is worth noting that the decomposition of PF is accompanied by the generation of H2 O which will lead to a serials side reaction and make the purification more difficult. From this perspective, TOX is more superior to PF. However, the reversible process of TOX decomposition is more facile by 5.92 kcal mol−1 than that of PF decomposition. Moreover, as IM1 illustrated in Fig. 1, the linear trioxymethylene complex yielded during TOX decomposition is stabilized by [HSO4 ]− via forming C O bond and is energetically comparable with the reaction entrance. So we suspect its formation will bring the loss of catalysts and TOX. Both these two factors suggesting TOX may be less efficient than PF on supplying FA when SO3 H-FILs are chosen as the catalysts. 3.2. The reaction of MeOH with FA The optimized geometries of intermediates and transition states are illustrated in Fig. 3. And the corresponding schematic energy surface is shown in Fig. 4. As can be seen in Fig. 3, a three-body complex IM7 is firstly formed from SO3 H-FILs, MeOH and FA via two O H· · ·O hydrogen bonds with a binding energy of 6.89 kcal mol−1 . Subsequently, pro-
231
tons transfer through these two hydrogen bonds via TS7–8 leading to the formation of methyl hemiformal IM8. The energy barrier to be overcome for this process is 11.45 kcal mol−1 , implying the formation of hemiacetal is very facile. Starting from IM8, four potential pathways have been located, and they are denoted as paths I, II, III, and IV, respectively. Along path I, with the introduction of MeOH, IM9 is firstly formed directly, which is 4.79 kcal mol−1 higher than that of IM8. Moderate hydrogen bond O H· · ·O between MeOH and cation holds IM8 and MeOH together. Subsequently, with the assistance of SO3 H-FILs, IM9 can go through TS9–10 to form DMM and liberate a H2 O molecule. The structure of TS9–10 shown in Fig. 3 illustrates the whole process: OH of methyl hemiformal abstracts the H atom of HSO4 − leading to the formation of H2 O and carbocation C1 + . Simultaneously, SO3 H group transfers its H atom to one O atom of SO4 2− and accepts another H atom from MeOH making the recovery of SO3 H-FILs catalyst. The formed carbocation reacting with methoxy makes the generation of DMM. The whole process occurs through a single step with a barrier of 21.45 kcal mol−1 (relative to zero point energy, which is applied to the barriers of the forward reactions discussed in the following sections if not pointed out particularly). For path II, in the presence of FA, weak van der Waals interaction between FA and methyl hemiformal in IM8 hold them together to form IM11. Subsequently, accompanied with the liberation of H2 O, FA is added into the newly formed carbocation CH3 O CH2 + (C1 + ) with an energy barrier of 27.35 kcal mol−1 , and resulting in the formation of chain increased carbocation CH3 O CH2 O CH2 + (C2 + ), as depicted by IM12 shown in Fig. 3. That is to say, FA addition and H2 O departure occurs via a concerted step. Next, with the introduction of MeOH, the C2 + is capped leading to the generation of DMM2 (see IM13 → TS13–14 → IM14 shown in Fig. 3). The energy barrier about this step is 27.31 kcal mol−1 , which is comparable to that of the FA addition step. It is worth noting that the barriers to be overcome for the reverse reaction about these two steps are 18.74 and 19.95 kcal mol−1 , respectively, which are all lower than those of the forward reaction, suggesting the reverse reaction occurs more facile, i.e., DMM2 is more easily to be decomposed than its formation, which can explain the experiment observations that DMM is always the predominant product in the reaction of MeOH or DMM with TOX [3,9]. Can the first concerted elementary process along path II occur stepwisely, i.e., the methyl hemiacetal is firstly dissociated into H2 O and C1 + , and then FA inserts into C1 + resulting in the production of C2 + ? To answer this question we examined another pathway (III), as the geometries IM15-IM18 shown in Fig. 3 and the energy surface illustrated in Fig. 4. The calculated results demonstrate the dehydration process occurs easily and the energy barrier required is only 20.85 kcal mol−1 which is energetically 6.50 kcal mol−1 lower than that of the first step of path II (27.35 kcal mol−1 ). Thus we conjecture methyl hemiformal favors to dehydration firstly and then to react with FA although the energy barrier to be surmounted for the following FA addition step (27.57 kcal mol−1 ) is comparative with that of the concerted pathway (27.35 kcal mol−1 ), i.e., path III rather than path II is preferred. The geometry of obtained carbocation C2 + (see IM18) is similar to that shown in IM12. Thus, the succeeded reaction that C2 + capped by MeOH will occur similarly as the process illustrated in IM13 → TS13–14 → IM14 and hence it isn’t investigated further. Additionally, according the results discussed above, we conjecture adding FA into IM12 or IM18 continually will gain the chain grown carbocations, and capping these carbocations will result in the formation of DMMn with different chain length. Beginning with IM8, the reaction proceeding along path IV with FA inserting into O H bond of methyl hemiformal to make the
232
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
Fig. 2. Energy surface for the dissociation of paraformaldehyde (PF) catalyzed by SO3 H-FILs and corresponding optimized structures of intermediates and transition states. The summary energy of free PF and SO3 H-FILs is taken as the zero point energy.
chain propagate has also been considered, as IM19 IM22 shown in Fig. 3 and the energy surface with black lines depicted in Fig. 4. Obviously, the insertion of FA into O H bond is completed in one step: H atom of OH in methyl hemiformal is abstracted by O atom of FA and simultaneously O atom of methyl hemiformal attacks C atom of FA. The second FA is inserted into hemiacetal in a similar manner. The energy barriers for the insertion of
these two FA are 42.29 and 44.36 kcal mol−1 , respectively, which are much higher than those of path II for the addition of FA into C O bond. Therefore, we conjecture the reaction along this pathway is negligible so the pathway will not be investigated further. As can be seen IM19–IM22 in Fig. 3, during the addition process of FA, SO3 H-FILs does not act as a catalytic center, and its vital role is to stabilize the intermediates and transition states via
Fig. 3. Optimized structures for the intermediates and transition states included in the reaction of FA with MeOH catalyzed by SO3 H-FILs.
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
G: kcal mol-1 path I: formation of DMM path II: insertion of FA into C-O path III: dissociation a H2O from hemiacetal path IV: insertion of FA into O-H
TS19-20 42.29
TS21-22 44.36
TS11-12 27.35
TS17-18 27.57 27.31 TS13-14
TS9-10 21.45
IM7 6.89 0.00
20.85 TS15-16
TS7-8 11.45 IM8 1.68
IM9 6.47 2.98 IM11
233
IM19 IM20 5.08 6.92 3.96 IM15
IM12 11.33 8.30 IM16 4.21 IM10
IM17 12.44 8.61 IM13
IM21 10.29
7.36 IM14
IM22 8.18 7.36 IM18
Fig. 4. Schematic energy surface for the reaction of MeOH with FA catalyzed by SO3 H-FILs.
O H· · ·O hydrogen bonds, which rationalized the high energy barriers demanded. On the basis of the results discussed above, it is obviously found that the formation of methyl hemiformal is a very key step in the generation of DMMn , because the following processes are all started with this complex. Actually, methyl hemiformal has also been detected as the key intermediate by transient isotopic experiments in MeOH oxidation to synthesis DMM [36]. Above results demonstrate once the methyl hemiformal formed, the reaction probably proceeds along four pathways. When comparing these four pathways, it is noticeable found that the formation of DMM along path I, which only requires 21.45 kcal mol−1 to conquer the energy barrier, is the most favorable pathway while path IV which has the highest barrier (44.36 kcal mol−1 ) to be overcome can be ignored. When comparing paths II and III, it is found the only difference refers to the liberation of H2 O and the addition of FA is concerted or stepwise. The calculated results have illustrated path III is more favorable. As for paths I and III, higher energy barrier of path III than path I (by 6.12 kcal mol−1 ) suggesting the formation of DMM2 is more difficult than DMM, which is in good accordance with the experimental observations [3,9]. Based on the discussion of the four pathways, we speculate the formation of DMMn from MeOH and FA will occur along the following hemiacetal-carbocation mechanism: MeOH firstly reacts with FA leading to the formation of methyl hemiformal, and then methyl hemiformal can react with MeOH leading to the formation of DMM or dehydrate to generate C1 + carbocation which can be stabilized by ILs. If introducing FA, the carbocation chains will grow and if providing MeOH the carbocation chains can be capped resulting in the formation of DMMn with different chain lengths. During these processes, SO3 H-FILs play crucial roles in every elementary step. On the one hand, via protons transfer, the cation and the anion promote the formation and the dehydration of methyl hemiformal, and the formation, the growth and the termination of the carbocations. On the other hand, the formed carbocations can be stabilized by O atom of cation. In a word, the cations and anions of SO3 H-FILs cooperatively promote the reaction of MeOH with FA resulting in the formation of DMMn with different chain length. It is worth noting that every process related to the reaction is reversible, suggesting the reaction of MeOH with FA is in equilibrium which is in good agreement with the investigation of Burger et al. [9,22,37,38] and the suggestion of Arvidson [39]. Additionally, as can be seen in Fig. 4, all the intermediates and transition states are energetically higher than the initial reactants, meaning that the
reactions should be carried out under mild heating conditions. A side product H2 O, which is unfavorable for the formation of DMMn , is generated during the reaction of MeOH with FA, indicating MeOH is probably not a good end-capping provider [24]. 3.3. The reaction of DMM with FA According to the results discussed above, the reaction of MeOH with FA follows a hemiacetal-carbocation mechanism. To clarify if this mechanism also applies to the reaction of DMM with FA mediated by SO3 H-FILs, a further study has been conducted. Fig. 5 lists the optimized species involved in the reaction. And Fig. 6 shows the corresponding energy surface. As can be seen in Fig. 5, the reaction starts with the formation of complex IM23 from DMM and SO3 H-FILs via a strong O H· · ·O hydrogen bond which lies 7.81 kcal mol−1 above the reaction entrance. Then, IM23 is converted into IM24 via TS23–24 with a barrier of 25.33 kcal mol−1 . In TS23–24 , the H atom of SO3 H group is migrating to O1 atom of DMM along the initially formed hydrogen bond making the C1 O1 bond be elongated from 1.450 Å to 2.246 Å. Simultaneously, the distance between C1 and O2 atoms are reducing from 3.116 Å to 2.206 Å. The transition vector indicated by vibration analysis as well as the subsequent IRC calculations manifest the formation of MeOH and carbocation C1 + which is stabilized by HSO4 − (see IM24). Next, FA gradually approaches C1 + leading to the chain increased carbocation C2 + which is stabilized by O atom of [BsMim]+ cation, as depicted by the process of IM25 → TS25-26 → IM26 shown in Fig. 5. The overall barrier for this step is 28.15 kcal mol−1 . The following step of the reaction refers to an end capping process of IM26 by MeOH forming DMM2 with TS27–28 as the transition state. And the barrier to be surmounted is 10.75 kcal mol−1 . However, if FA is presented, it may react with IM26 resulting in the formation of chain length increased carbocation C3 + with a barrier of 17.42 kcal mol−1 , as the process IM29 → TS29–30 → IM30 illustrated in Figs. 5 and 6. C3 + can be terminated with a lower energy barrier 6.94 kcal mol−1 if MeOH is introduced, as depicted by IM30 → TS30–31 → IM31. The chain length of C3 + will continually increase if keeping on introducing FA. Above results clearly indicate the formation of DMMn from DMM and FA is carried out via a carbocation mechanism: DMM is protonated to generate carbocation C1 + firstly, then adding FA one by one to C1 + will make the carbocation chain increase, and end-capping the formed carbocations by MeOH will result in the generation of DMMn with different chain length. For clarity, the formation process of DMMn derived from DMM and FA are sum-
234
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
Fig. 5. Optimized structures for the intermediates and transition states included in the reaction of DMM with FA catalyzed by SO3 H-FILs.
Fig. 6. Energy surface for the reaction of DMM with FA catalyzed by SO3 H-FILs.
Scheme 3. The formation process of DMMn from DMM and FA.
marized in Scheme 3. According to these results, it is probably to conclude that the formation of DMMn (n ≥ 2) is easily to achieve. However, for the same carbocation, the selectivity for FA addition and MeOH end-capping processes is different. The barriers to be surmounted for end-capping processes by MeOH to form DMM2 and DMM3 (10.75 and 6.94 kcal mol−1 for process IM26 → IM27 → TS27–28 → IM28 and IM30 → TS30–31 → IM31, respectively) are always lower than those of FA addition processes (17.42 kcal mol−1 for process IM29 → TS29–30 → IM30 and
17.06 kcal mol−1 for process IM32 → TS32–33 → IM33), suggesting the end-capping process by CH3 OH is prior to the FA addition step, which is in good agreement with the investigation of Zhao et al. [23]. Moreover, the reverse reactions for FA addition processes are more and more favorable than their forward process with the chain increase, further demonstrating the propagation of carbocation is more and more difficult. When comparing the energies of the carbocations with different lengths, such as IM26, IM30, and IM33, it is noticeably found that the longer the chain the less sta-
0.10 0.08 0.06 0.04 0.02 0.00 DMM1 7h DMM2 6h DMM4 5h DMM5 4h DMM6 3h DMM7 2h DMM8 1h DMM9 20min DMM10 DMM11 0min
Molar distribution (mol/g)
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
Fig. 7. DMM1–2,4–11 molar distributions resulting from the decomposition of DMM3 promoted by SO3 H-FILs at 30 ◦ C and 0.1 MPa. The reaction time is 7 h.
ble the carbocation. All of these factors, i.e., high selectivity for end-capping process, less favorable addition procedure of FA, and reduced stability of carbocation with long chains, contribute to the lower production of DMMn with longer chains. These results can be used to understand the experimental observations that the production of DMMn is always lower than that of DMMn−1 . In addition, we find another interesting phenomenon that the barriers about the reverse process of every elementary step are always lower or comparable with the forward processes, indicating the reverse reactions are energetically more favorable than the forward reactions. For example, the calculated results indicate the dissociation of DMM2 to regenerate DMM only requires 27.21 kcal mol−1 to surmount the highest transition state TS27–28 , which is easier by 1.91 kcal mol−1 than its formation by DMM reacting with FA. This result is in good accordance with the experimental observations that the DMM content is always higher than that of DMM2 [3,9]. Similarly, the formation of DMM3 demands more energy to surmount the barriers than its decomposition. Hence, we speculate the temperature required for DMMn (n > 1) decomposition will be lower than that for its formation. To check out if the conclusions obtained above apply to the real reaction system, we have examined the decomposition of DMM3 mediated by SO3 H-FILs at 30 ◦ C and 0.1 MPa. The product molar distribution is shown in Fig. 7. It is interesting to state that DMMn with different chain lengths were observed at the initial stage of the reaction. Moreover, the content of DMM is the highest, while DMM2,4–11 decrease gradually. However, the reaction of DMM with TOX or PF catalyzed by SO3 H-FILs hardly occurs under same conditions. It should be carried out at least 80 ◦ C and 0.1 MPa. All of these results confirm the calculated results that the decomposition of DMMn (n > 1) is easier to carry out than its formation. In addition, the calculated results that the formation of DMMn with longer chain is easy to carry out in spite of different preference have also been validated by present experimental results. Based on the calculated results, during the decomposition of DMM3 , the formation of IM26 is the most demanding step which requires 26.06 kcal mol−1 to surmount the energy barrier, as the backward process DMM3 + IL → IM26 illustrated in Fig. 6. Once this step is passed through successfully, the following steps started with IM26 which refer to the formation of DMM1,2 would proceed rapidly. Especially, the formation of DMM is energetically more favorable by 0.97 kcal mol−1 than that of DMM2 , indicating DMM is preferred to generate to DMM2 . Thus, no overshoot in the con-
235
centrations of DMM2 will appear. For the formation process of DMM4 , it is found that the generation of carbocation C4 + via TS32–33 is isoenergetic with the FA release process of DMM3 (see IM30 → TS29–30 → IM29). In combined with the more favorable backward reaction of C4 + generation process, we speculate there also no overshoot exists in the concentrations of DMM4 . All of these results are consistent with the experimental findings shown in Fig. 7. Another interesting question refers to the formation of DMMn with different lengths at the same time or in sequence? Burger et al. [9] and Zheng et al. [10] have discussed this question in their reports, but contrast conclusions were obtained. Burger et al. thought DMMn were formed simultaneously while Zheng et al. considered DMMn were synthesized in sequence. Our present experimental results which were shown in Fig. 7 indicated that DMM1–2,4–10 were simultaneously detected at the initial stage. Our calculated results demonstrate once the decomposition of DMM3 to form carbocation C2 + (DMM3 + IL → IM26) can be achieved successfully the other steps would occur rapidly in spite of a different preference (see Fig. 6). Additionally, the way for the carbocation chain propagation and the termination process of these chains indicated that the formation of DMMn was not derived from DMMn−1 except the dissociation cases of DMMn−1 resulting in the formation of DMMn . In other words, all DMMn (n ≥ 2) could be produced directly from DMM via FA insertion rather than forming DMMn−1 firstly and then yielding DMMn from DMMn−1 (see Scheme 3). Based on these results, we conjecture the formations of DMMn are probably at the same time. We hope this question be further elucidated in our future kinetic studies. During the whole process, SO3 H-FILs play important roles. Firstly, the H atom of SO3 H-FILs is “borrowed” by DMM to release a MeOH and a carbocation, and then the formed carbocation can be stabilized by O atom of SO3 H-FILs. At last, the H atom is “returned” to SO3 group during the termination process making the regeneration of SO3 H-FILs. These results indicate both cations and anions affect the catalytic activity of ionic liquids on the formation of DMMn . Additionally, it has been confirmed that many protic ionic liquids possess brönsted acidic nature, such as [HMIM]+ [HSO4 ]− [40]. So it will be valuable to investigate the effects aroused by different cations, anions and protic ionic liquids on the formation of DMMn . However, this is another issue which beyond the scope of present work and will be pursued in our future work. 4. Conclusions The detailed formation mechanism of DMMn from MeOH (DMM) with TOX or with PF promoted by SO3 H-FILs has been examined using DFT calculations. The following conclusions can be obtained: (1) Before reacting with MeOH or DMM, TOX and PF will decompose into monomeric FA firstly. The decomposition of TOX follows a two-step mechanism: TOX protonated by SO3 H-FILs firstly leading to the formation of the ring-opened trioxymethylene chain, and then the transferred hydrogen atom goes back to SO3 H-FILs realizing the formation of three FA monomers. As for PF, its decomposition to produce monomeric FA is carried out by one-step mechanism. (2) When MeOH is chosen as the capping provider, the condensation process follows a hemiacetal-carbocation mechanism: MeOH reacting with FA results in the formation of hemiacetal which is a key intermediate for the reaction. This hemiacetal can react with MeOH leading to the generation of DMM or react with FA to bring out the production of chain increased carbocations. Capping these carbocations by MeOH will obtain
236
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
the DMMn . While DMM is selected to terminate the end, a carbocation mechanism will be preferred. With the assistant of SO3 H-FILs, DMM firstly dissociates into MeOH and a carbocation. Inserting FA into this carbocation one by one will generate the chain grown carbocations which can be terminated by MeOH leading to the formation of DMMn . (3) The product distribution of DMMn that DMMn < DMMn−1 can be understood by the following points: 1) for the same carbocation, its termination by MeOH is always more favorable than its chain growth process by the addition of FA; 2) the reverse reaction of FA addition step proceeds more rapidly than the forward reaction which brings difficulty for the carbocation growth; 3) the stability of carbocations reduces with the increase of chain lengths; 4) the formation of DMMn is more difficult than its decomposition. (4) SO3 H-FILs play vital roles no matter in the decomposition of TOX/PF or in the reaction of FA with MeOH/DMM. On the one hand, the cations and anions of SO3 H-FILs can synergistically promote the reaction via protons transfer. On the other hand, the formed carbocation can be stabilized by SO3 H-FILs via forming C O bonds. In addition, the intermediates and transition states can also be stabilized through their hydrogen-bonding interactions with SO3 H-FILs. (5) Both the decomposition of TOX/PF and the reaction of FA with MeOH/DMM are equilibrium process. Moreover, the reactions should be operated under heating conditions.
[3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25]
Acknowledgements This work was supported by the National Basic Research Program of China (973 Program, No. 2011CB201404), and National Natural Science Foundation of China (Project No. 21203218, 21133011, 21473225). We gratefully acknowledge the computing resources and time on the Supercomputing Center of Cold and Arid Region Environment and Engineering Research Institute of Chinese Academy of Sciences. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.molcata.2015.07. 028 References [1] F. Willian, E. Richard, US: 2449469 (1948). [2] F. Zhao, H.J. Li, H.L. Song, C.G. Xia, Nat. Gas Chem. Ind. 1 (2013) 1–6.
[26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40]
H.J. Li, H.L. Song, L.W. Chen, C.G. Xia, Appl. Catal. B 165 (2015) 466–476. P.H. Gary, J.S., Michael, US: 6160174A (2000). P.H. Gary, J.S., Michael, US: 2003/0171534A1 (2003). P.H. Gary, J.S., Michael, US: 6437195 B2 (2002). L. Wang, W.T. Wu, T. Chen, Q. Chen, M.Y. He, Chem. Eng. Commun. 201 (2014) 709–717. T. Chen, L. Wang, Q. Chen, M.Y. He, Ion Exc. Ads. 28 (2012) 456–462. ¨ J. Burger, E. Strofer, H. Hasse, Ind. Eng. Chem. Res. 51 (2012) 12751–12761. Y.Y. Zheng, Q. Tang, T.F. Wang, Y.H. Liao, J.F. Wang, Chem. Eng. Technol. 36 (2013) 1951–1956. Q. Zhao, H. Wang, Z.F. Qin, Z.W. Wu, J.B. Wu, W.B. Fan, J.G. Wang, J. Fuel Chem. Technol. 39 (2011) 918–923. X.H. Zhang, P.T. Wu, Y. Zhang, Z.C. Jiang, T.T. Liu, X.Y. Gao, Y.Q. Hu, Chem. React. Eng. Technol. 30 (2014) 140–144. J.Q. Zhang, D.Y. Fang, D.H. Liu, Ind. Eng. Chem. Res. 53 (2014) 13589–13597. J. Chen, Z.H. Tang, C.G. Xia, X.Z. Zhang, Z. Li, CN: 101182367A (2008). J. Chen, H.Y. Song, C.G. Xia, X.Z. Zhang, Z.H. Tang, CN: 101665414A (2010). J. Zhang, M. Shi, D. Fang, D. Liu, React. Kinet. Mech. Catal. 113 (2014) 459–470. H.J. Li, H.L. Song, F. Zhao, L.W. Chen, C.G. Xia, J. Energy Chem. 24 (2015) 239–244. J. Chen, H.Y. Song, C.G. Xia, Z. Li, CN: 102249869A (2011). R. Patrini, M. Marchionna, EP: 1505049A1 (2005). R. Patrini, M. Marchionna, US: 6534685 (2003). H. Schelling, E. Ströfer, R. Pinkos, A. Haunert, G. Tebben, H. Hasse, S. Blagov, WO: 2006045506 A1, (2006). E. Ströfer, H. Hasse, S. Blagov, US: 20080207954 A1 (2008). Y.P. Zhao, Z. Xu, H. Chen, Y.C. Fu, J.Y. Shen, J. Energy Chem. 22 (2013) 833–836. J. Burger, M. Siegert, E. Ströfer, H. Hasse, Fuel 89 (2010) 3315–3319. M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery, J.E. Peralta Jr., F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, O. Farkas, J.B. Foresman, J.V. Ortiz, J. Cioslowski, D.J. Fox, Gaussian 09, Revision B.01, Gaussian, Inc., Wallingford, CT, 2010. S. Grimme, J. Comp. Chem. 27 (2006) 1787. W.J. Hehre, R. Ditchfield, J.A. Pople, J. Chem. Phys 56 (1972) 2257–2261. P.C. Hariharan, J.A. Pople, Theor. Chim. Acta 28 (1973) 213–222. G.W. Spitznagel, T. Clark, J. Chandrasekhar, P.v.R. Schleyer, J. Comput. Chem. 3 (1982) 363–371. T. Clark, J. Chandrasekhar, G.W. Spitznagel, P.v.R. Schleyer, J. Comput. Chem. 4 (1983) 294–301. C. Peng, H.B. Schlegel, Isr. J. Chem. 33 (1993) 449–454. H.B. Schlegel, J. Comput. Chem. 3 (1970) 214. K. Fukui, J. Phys. Chem. 74 (1970) 4161–4163. H.Y. Song, Z. Li, J. Chen, C.G. Xia, Catal. Lett. 142 (2012) 81–86. A.C. Cole, J.L. Jensen, I. Ntai, K.L.T. Tran, K.J. Weaver, D.C. Forbes, J.H. Davis, J. Am. Chem. Soc. 124 (2002) 5962. J.M. Tatibouët, L.P. Hélène, J. Mol. Catal. A 171 (2001) 205–216. J. Burger, H. Hasse, Chem. Eng. Sci. 99 (2013) 118–126. J. Burger, E. Ströfer, H. Hasse, Chem. Eng. Res. Des. 91 (2013) 2648–2662. M. Arvidson, M.E. Fakley, M.S. Spencer, J. Mol. Catal. 41 (1987) 391–393. T.L. Greaves, C.J. Drummond, Chem. Rev. 108 (2008) 206–237.
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical journal homepage: www.elsevier.com/locate/molcata
Mechanistic study for the formation of polyoxymethylene dimethyl ethers promoted by sulfonic acid-functionalized ionic liquids Fang Wang, Gangli Zhu, Zhen Li, Feng Zhao, Chungu Xia ∗ , Jing Chen State Key Laboratory for Oxo Synthesis and Selective Oxidation, Suzhou Research Institute of LICP, Lanzhou Institute of Chemical Physics (LICP), Chinese Academy of Sciences, Lanzhou 730000, PR China
a r t i c l e
i n f o
Article history: Received 5 February 2015 Received in revised form 14 July 2015 Accepted 24 July 2015 Available online 31 July 2015 Keywords: Polyoxymethylene dimethyl ethers Sulfonic acid-functionalized ionic liquids Mechanism DFT
a b s t r a c t Polyoxymethylene dimethyl ethers (DMMn ), which are ideal additives for diesel fuel, are mainly synthesized from the condensation of methanol (MeOH) or dimethoxymethane (DMM) with 1,3,5-trioxane (TOX) or paraformaldehyde (PF) promoted by different acid catalysts. However, up to date, few studies have been reported to examine the formation mechanism of DMMn which is essential in understanding the reaction and valuable in designing improved catalysts. In this work, using the density functional theory (DFT) calculations combined with experiment studies, we evaluate the formation mechanism of DMMn which is promoted by sulfonic acid-functionalized ionic liquids (SO3 H-FILs). Our calculated results indicate TOX and PF should dissociate into formaldehyde monomers firstly and then to react with MeOH or DMM. However, their decomposition process is different where the dissociation of TOX proceeds along a two-step mechanism while it follows a one-step mechanism for PF dissociation. As for the formation of DMMn , the reaction proceeds along a hemiacetal-carbocation pathway when MeOH is selected as the capping group provider, while the reaction follows a carbocation pathway when DMM is chosen. The origination for the product distribution pattern has also been discussed in detail. The cations and anions of ionic liquids are found synergistically promote the condensation reaction by proton transfer and simultaneously stabilizing the formed intermediated and transition states. Moreover, all the processes related to the decomposition of TOX and PF and the condensation reaction are reversible. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Polyoxymethylene dimethyl ethers (DMMn , generally 1 ≤ n ≤ 8) are a series chain compounds with the formula CH3 O(CH2 O)n CH3 containing high oxygen contents (42–51%) and cetane numbers (CN). Especially, the CN of DMMn with n = 3–8 is up to 76 or more and the oxygen contents is between 47% and 50%, making them be considered as the promising green additives to diesel fuel, which can lead to the improved thermal efficiency, and the significantly reduced emissions of particular materials, such as COx and NOx . Therefore, the synthesis and application of DMMn have attracted worldwide attentions [1–13]. Two type raw materials are required for the synthesis of DMMn : one refers to a compound providing CH2 O group as a chain segment, such as 1,3,5-trioxane (TOX), paraformaldehyde (PF), or formaldehyde (FA), and the other is a compound which offers CH3 -capping group, such as methanol (MeOH), dimethoxymethane (DMM), or dimethyl ether (DME).
∗ Corresponding author. Fax: +86 931 4968129. E-mail address: [email protected] (C. Xia). http://dx.doi.org/10.1016/j.molcata.2015.07.028 1381-1169/© 2015 Elsevier B.V. All rights reserved.
Both these two type compounds can be synthesized from MeOH which mainly produced from coal and natural gas and facing seriously overplus. Thus, the synthesis of DMMn also promotes the transformation of MeOH into clean energy. So far, many catalysts have been developed for the condensation reaction, such as liquid acids [1], solid superacid [2,3], ion exchange resin [4–10], molecular sieves [11,12], metal oxide [13] and so on. However, few catalysts are suitable for industrial application, so the exploration of effective and environmentally benign catalysts for the acetalization process needs to be solved urgently as oil resources are dwindling gradually. In recent years, our group have developed several sulfonic acid-functionalized ionic liquids (SO3 H-FILs) which are found very effective in promoting the condensation reaction [14,15]. The consecutive research found this catalyst possess many advantages, such as enhance the conversion of raw materials up to 90%, increase the selectivity for DMM3–8 up to 40% and so on, which suggest the condensation reaction catalyzed by SO3 H-FILs exhibiting promising application prospect. For the formation of DMMn , there are several key points we concerned: (1) do TOX and PF decompose firstly before they reacting with MeOH or DMMn ? (2) how does the chain length actually
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
propagate? (3) do different capping group providers, such as MeOH and DMM, lead to distinct reaction pathways? (4) what roles do the catalysts indeed play in the reaction? However, to the best of our knowledge, the published articles and patents are concentrated on the synthesis of DMMn , the development of catalysts [3,4,10], reaction kinetics [9,13,16,17], and chemical process design [18–22]. Few computational studies have been conducted to investigate the questions listed above to date at a molecular level for this important reaction regardless of which kind catalysts being used. In particular, employing experimental methods Zhao et al. [23] have found the growth of DMMn chain is carried out via addition FA one by one and the products distribution follow the Schulz–Flory law which restricts the increase of the selectivity to the desired products. However, the detailed reaction pathways are still unknown. Zheng et al. based on their experiment studies proposed DMMn + 1 are probably formed from the directly addition of protonated FA into C O bond of DMMn [10]. However, the proposed mechanism has not been validated no matter by computational analysis or experimental methods. As for the possible form of TOX or PF to react with end-capping species, Burger et al. have proposed TOX may decompose into a linear trioxymethylene chain firstly and then this chain complex directly inserted into another DMMn leading to the formation of DMMn+3 , such as DMM4 , DMM7 and so on. But they didn’t observe the favorite formation of DMM4 and DMM7 in experiments [9,24]. To clarify the problems listed above, we have carried out a detailed density functional theory (DFT) study for the reactions of TOX/PF with MeOH and with DMM, as depicted in Scheme 1. Based on previous reports of our group [14], 1-(4-sulfonic acid)-butyl-3methylimidazolium hydrogen sulfate [BsMim][HSO4 ] are chosen as the catalysts. We expect the present results will provide new insights into the formation mechanism of DMMn and explain the previous experimental observations in some extent.
229
equipped with a flame ion-ization detector (FID) and a SE-54 capillary column (60 m × 0.25 mm × 0.25 m). High purity nitrogen was used as the carrier gas. It is worth noting that, to avoid further process of the reaction, the ionic liquids have been separated from the sample before GC analysis. Firstly, acetic ether (2.0 g), internal standard (furanidine, 0.02 g), and small amounts CaO (which was used to neutralize any possible acids in the sample mixture) were put into a vial. Then the reaction sample (about 0.2 g) was rapidly transferred to the vial. Thus, ionic liquids were separated owing to its insolubility in acetic ether. The unconverted FA is ascertained by titration method. Two drops of thymol phthalein indicator were added into the sodium sulfite solution (1 mol/L, 20.0 g) firstly, and then this solution was neutralized to colorless by sulphuric acid. Next, the reaction sample (1.0 g) was added into the obtained solution and shocked to blue. At the end, the content of FA can be determined by the titration of sulphuric acid (0.1 mol/L). Molar distribution of every component was adopted to clearly show the content of DMMn yielded from the decomposition of DMM3 , which is defined as md (n, t) =
mDMMn /MDMMn mtotal
where mtotal = the sum mass of all components which is set as 100 g, mDMMn = the mass of DMMn , MDMMn = molar mass of DMMn . The unit of molar distribution is mol/100 g. During the decomposition of DMM3 , the total mass of all components is a constant while the total mole amounts of DMMn keep varying. To clearly reflect the changes of the individual component, the total mass of DMMn was chosen as the denominator. It is worth noting that the molar distribution (mol/g) is not given as the complement to 100% while the total mass fraction (g/g) would be 100%. 3. Results and discussion
2. Computational and experimental details 3.1. The decomposition of TOX and PF 2.1. Computational methods All the DFT calculations were carried out with the Gaussian 09 software package [25]. The hybrid GGA functional B97D [26] which includes dispersion corrections combined with 6-31+G(d,p) [27–30] basis set was selected for the geometry optimization. All geometries were fully optimized without any symmetry constraints. The geometries for the transition states were located employing the synchronous transit-guided quasi-Newton (STQN) method [31] and Berny algorithm [32]. Vibrational frequency analysis was performed to confirm the nature (minima or first-order saddle points) of the stationary points at the same level. Zero-point energy corrections were carried out for all calculated energies. The intrinsic reaction coordinate (IRC) [33] was conducted in both directions (forward and reverse) from the transition states to the corresponding local minima to identify the minimum-energy paths. To take the entropy effects into account, the Gibbs free energies (G) were used in the following discussion. 2.2. Experimental details [BsMim][HSO4 ] were prepared according to the similar procedure reported previously [34,35]. DMM3 (99.9%, technical) was separated from DMMn mixtures. For the decomposition of DMM3 , DMM3 and [BsMim][HSO4 ] catalyst with a mass ratio of 50:1 were charged into a 100 mL glass reactor. The reaction was performed with rapid agitation by magnetic stirrer at 30 ◦ C and 0.1 MPa. The products were identified and quantitatively analyzed by gas chromatography (GC, Agilent 6890)
Before analyzing the reaction mechanism, the chain segment provider TOX and PF dissociate or not with the assistance of SO3 HFILs were firstly examined. This will be helpful to understand how DMMn chains increase in the following steps. 3.1.1. TOX decomposition As can be seen in Fig. 1, beginning with the SO3 H-FIL + TOX, a two-body complex IM1 is firstly formed via hydrogen-bonding interaction between H atom of SO3 H moiety and one O atom of TOX. Subsequently, IM1 can go through TS1–2 to open the six-membered ring and produce a linear trioxymethylene chain complex where the carbocation side is stabilized by the O atom of HSO4 − anion (see IM2). This process is carried out via the migration of H atom of SO3 H to O atom of TOX making one C O bond of TOX be broken. 20.22 kcal mol−1 is required to conquer the energy barrier. It is worth to note that the geometry of trioxymethylene chain is very similar to that of hemiacetal. And then the transferred hydrogen atom goes back to SO3 H-FIL making the decomposition of TOX, i.e., the formation of monomeric FA, as the large C O bond distances 2.072 and 1.968 Å have been illustrated in IM4. The saddle point about this step is TS3–4 with a barrier of 13.45 kcal mol−1 . The separation of monomeric FA from SO3 H-FIL, which is exothermic by 15.18 kcal mol−1 , completes the whole process. According to the results discussed above, it is obviously found that the decomposition of TOX into FA monomers occurs via two steps: (i) the protonation of TOX by SO3 H-FILs resulting in the six-membered ring opening and the formation of a linear trioxymethylene chain complex, (ii) the decomposition of the chain complex to produce
230
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
Scheme 1. The reaction of TOX/PF with CH3 OH and DMM.
the monomeric FA. The rate-determining step refers to the first step, i.e., the protonation of TOX. Additionally, taking free SO3 H-FILs and monomeric FA as the starting point, the reverse reaction can also occur. It overcomes a relatively higher barrier 22.24 kcal mol−1 via TS3–4 forming a trioxymethylene chain complex and then surmounts a little lower energy barrier 20.16 kcal mol−1 via TS1–2 leading to the regeneration of TOX. Apparently, the first step which is more energy demanding is the rate-determining step for the reverse reaction. A little higher energy barrier by 2.02 kcal mol−1 compared with the forward reaction suggests although the decomposition of TOX is reversible, the forward reaction, i.e., the formation of monomeric FA, is more favorable. Another point we interested is if the linear trioxymethylene chain can further react with MeOH or DMM directly leading to the formation of DMM4 . To clarify this question, the corresponding mechanistic details have also been investigated. We first explored
the capping process of linear trioxymethylene by MeOH (see Fig. S1 in Supporting information). The calculated results indicate the energy should be surmounted is as high as 49.47 kcal mol−1 , indicating this process is difficult to realize in mild heating conditions. For the reaction of trioxymethylene chain with DMM, the calculated energy barrier is 51.53 kcal mol−1 , demonstrating this process is also not feasible. Both these results suggest the formation of DMM4 via trioxymethylene chain reacting with MeOH or DMMn directly will be negligible, which is also not observed in experiment [24]. Based on the results discussed above, it is noticeably found that SO3 H-FILs plays important roles in the decomposition process of TOX. On the one hand, hydrogen transfer along hydrogen bonds between cation and TOX can promote the ring opening and the decomposition of TOX. On the other hand, the formed carbocation trioxymethylene chain can be stabilized by O atoms of HSO4 − anion. In other words, the cation and anion can synergistically promote the TOX decomposition.
Fig. 1. Energy surface for the decomposition of 1,3,5-trioxane (TOX) catalyzed by SO3 H-FILs and corresponding optimized structures of intermediates and transition states. The summary energy of free TOX and SO3 H-FILs is taken as the zero point energy. Distances are in Å and energies are in kcal mol−1 . For clarity, the structures of imidazolium and alkyl side chain are depicted in line.
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
Scheme 2. PF decomposition.
3.1.2. PF decomposition After clarifying the decomposition mechanism of TOX, our attentions turns to the dissociation of PF. Considering the calculation amount, the degree of PF polymerization we chosen is 3. Being different from the structure of TOX, the ends of PF contain an H atom and an OH group, respectively, which make the decomposition of PF liberate a H2 O molecule, as depicted in Scheme 2. At the reaction entrance, PF binding to SO3 H-FILs via strong hydrogenbonding interactions results in the formation of IM5 (see Fig. 2), which is energetically below the entrance by 5.96 kcal mol−1 . Subsequently, H1 atom of SO3 H group transfers to the oxygen atom of one OH group resulting in the liberation of H2 O, as depicted by TS5–6 in Fig. 2. Simultaneously, H2 atom from the other OH group of PF transfers to HSO4 − anion, and H3 atom of HSO4 − anion shifts to one oxygen atom of SO3 group. All of these processes contribute to the one-step decomposition of PF and the regeneration of SO3 HFILs catalyst. The energy barrier of this step is 23.89 kcal mol−1 , which is easily to be overcome under mild heating condition. The moderate hydrogen-bonding interaction between monomeric FA and SO3 H-FILs make the catalyst dissociation very facile, which is exothermic by 12.52 kcal mol−1 . The barrier for the backward reaction is 28.16 kcal mol−1 , which is higher than the forward process by 4.27 kcal mol−1 , implying the regeneration of PF is energetically less favorable. However, the barriers 23.89 and 28.16 kcal mol−1 is not too high to be overcome under heating conditions. Thus, the decomposition of PF is also an equilibrium reaction. Additionally, from the geometries of TS5–6 , it is obviously found that SO3 H-FILs play indispensable roles for PF decomposition via protons transfer. The liberation of SO3 H-FILs accomplished the whole process. From the discussion given above, it is clear that the dissociation of TOX proceeds along a two-step pathway, while it is realized via one step for PF decomposition. Moreover, both the decomposition of TOX and PF is facile to proceed under acid existing and mild heating conditions, suggesting they may firstly dissociate into FA monomers and then to react with MeOH or DMM. Thus, FA instead of TOX and PF is used directly in the following discussions about the mechanistic details of the DMMn formation process. In addition, it is worth noting that the decomposition of PF is accompanied by the generation of H2 O which will lead to a serials side reaction and make the purification more difficult. From this perspective, TOX is more superior to PF. However, the reversible process of TOX decomposition is more facile by 5.92 kcal mol−1 than that of PF decomposition. Moreover, as IM1 illustrated in Fig. 1, the linear trioxymethylene complex yielded during TOX decomposition is stabilized by [HSO4 ]− via forming C O bond and is energetically comparable with the reaction entrance. So we suspect its formation will bring the loss of catalysts and TOX. Both these two factors suggesting TOX may be less efficient than PF on supplying FA when SO3 H-FILs are chosen as the catalysts. 3.2. The reaction of MeOH with FA The optimized geometries of intermediates and transition states are illustrated in Fig. 3. And the corresponding schematic energy surface is shown in Fig. 4. As can be seen in Fig. 3, a three-body complex IM7 is firstly formed from SO3 H-FILs, MeOH and FA via two O H· · ·O hydrogen bonds with a binding energy of 6.89 kcal mol−1 . Subsequently, pro-
231
tons transfer through these two hydrogen bonds via TS7–8 leading to the formation of methyl hemiformal IM8. The energy barrier to be overcome for this process is 11.45 kcal mol−1 , implying the formation of hemiacetal is very facile. Starting from IM8, four potential pathways have been located, and they are denoted as paths I, II, III, and IV, respectively. Along path I, with the introduction of MeOH, IM9 is firstly formed directly, which is 4.79 kcal mol−1 higher than that of IM8. Moderate hydrogen bond O H· · ·O between MeOH and cation holds IM8 and MeOH together. Subsequently, with the assistance of SO3 H-FILs, IM9 can go through TS9–10 to form DMM and liberate a H2 O molecule. The structure of TS9–10 shown in Fig. 3 illustrates the whole process: OH of methyl hemiformal abstracts the H atom of HSO4 − leading to the formation of H2 O and carbocation C1 + . Simultaneously, SO3 H group transfers its H atom to one O atom of SO4 2− and accepts another H atom from MeOH making the recovery of SO3 H-FILs catalyst. The formed carbocation reacting with methoxy makes the generation of DMM. The whole process occurs through a single step with a barrier of 21.45 kcal mol−1 (relative to zero point energy, which is applied to the barriers of the forward reactions discussed in the following sections if not pointed out particularly). For path II, in the presence of FA, weak van der Waals interaction between FA and methyl hemiformal in IM8 hold them together to form IM11. Subsequently, accompanied with the liberation of H2 O, FA is added into the newly formed carbocation CH3 O CH2 + (C1 + ) with an energy barrier of 27.35 kcal mol−1 , and resulting in the formation of chain increased carbocation CH3 O CH2 O CH2 + (C2 + ), as depicted by IM12 shown in Fig. 3. That is to say, FA addition and H2 O departure occurs via a concerted step. Next, with the introduction of MeOH, the C2 + is capped leading to the generation of DMM2 (see IM13 → TS13–14 → IM14 shown in Fig. 3). The energy barrier about this step is 27.31 kcal mol−1 , which is comparable to that of the FA addition step. It is worth noting that the barriers to be overcome for the reverse reaction about these two steps are 18.74 and 19.95 kcal mol−1 , respectively, which are all lower than those of the forward reaction, suggesting the reverse reaction occurs more facile, i.e., DMM2 is more easily to be decomposed than its formation, which can explain the experiment observations that DMM is always the predominant product in the reaction of MeOH or DMM with TOX [3,9]. Can the first concerted elementary process along path II occur stepwisely, i.e., the methyl hemiacetal is firstly dissociated into H2 O and C1 + , and then FA inserts into C1 + resulting in the production of C2 + ? To answer this question we examined another pathway (III), as the geometries IM15-IM18 shown in Fig. 3 and the energy surface illustrated in Fig. 4. The calculated results demonstrate the dehydration process occurs easily and the energy barrier required is only 20.85 kcal mol−1 which is energetically 6.50 kcal mol−1 lower than that of the first step of path II (27.35 kcal mol−1 ). Thus we conjecture methyl hemiformal favors to dehydration firstly and then to react with FA although the energy barrier to be surmounted for the following FA addition step (27.57 kcal mol−1 ) is comparative with that of the concerted pathway (27.35 kcal mol−1 ), i.e., path III rather than path II is preferred. The geometry of obtained carbocation C2 + (see IM18) is similar to that shown in IM12. Thus, the succeeded reaction that C2 + capped by MeOH will occur similarly as the process illustrated in IM13 → TS13–14 → IM14 and hence it isn’t investigated further. Additionally, according the results discussed above, we conjecture adding FA into IM12 or IM18 continually will gain the chain grown carbocations, and capping these carbocations will result in the formation of DMMn with different chain length. Beginning with IM8, the reaction proceeding along path IV with FA inserting into O H bond of methyl hemiformal to make the
232
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
Fig. 2. Energy surface for the dissociation of paraformaldehyde (PF) catalyzed by SO3 H-FILs and corresponding optimized structures of intermediates and transition states. The summary energy of free PF and SO3 H-FILs is taken as the zero point energy.
chain propagate has also been considered, as IM19 IM22 shown in Fig. 3 and the energy surface with black lines depicted in Fig. 4. Obviously, the insertion of FA into O H bond is completed in one step: H atom of OH in methyl hemiformal is abstracted by O atom of FA and simultaneously O atom of methyl hemiformal attacks C atom of FA. The second FA is inserted into hemiacetal in a similar manner. The energy barriers for the insertion of
these two FA are 42.29 and 44.36 kcal mol−1 , respectively, which are much higher than those of path II for the addition of FA into C O bond. Therefore, we conjecture the reaction along this pathway is negligible so the pathway will not be investigated further. As can be seen IM19–IM22 in Fig. 3, during the addition process of FA, SO3 H-FILs does not act as a catalytic center, and its vital role is to stabilize the intermediates and transition states via
Fig. 3. Optimized structures for the intermediates and transition states included in the reaction of FA with MeOH catalyzed by SO3 H-FILs.
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
G: kcal mol-1 path I: formation of DMM path II: insertion of FA into C-O path III: dissociation a H2O from hemiacetal path IV: insertion of FA into O-H
TS19-20 42.29
TS21-22 44.36
TS11-12 27.35
TS17-18 27.57 27.31 TS13-14
TS9-10 21.45
IM7 6.89 0.00
20.85 TS15-16
TS7-8 11.45 IM8 1.68
IM9 6.47 2.98 IM11
233
IM19 IM20 5.08 6.92 3.96 IM15
IM12 11.33 8.30 IM16 4.21 IM10
IM17 12.44 8.61 IM13
IM21 10.29
7.36 IM14
IM22 8.18 7.36 IM18
Fig. 4. Schematic energy surface for the reaction of MeOH with FA catalyzed by SO3 H-FILs.
O H· · ·O hydrogen bonds, which rationalized the high energy barriers demanded. On the basis of the results discussed above, it is obviously found that the formation of methyl hemiformal is a very key step in the generation of DMMn , because the following processes are all started with this complex. Actually, methyl hemiformal has also been detected as the key intermediate by transient isotopic experiments in MeOH oxidation to synthesis DMM [36]. Above results demonstrate once the methyl hemiformal formed, the reaction probably proceeds along four pathways. When comparing these four pathways, it is noticeable found that the formation of DMM along path I, which only requires 21.45 kcal mol−1 to conquer the energy barrier, is the most favorable pathway while path IV which has the highest barrier (44.36 kcal mol−1 ) to be overcome can be ignored. When comparing paths II and III, it is found the only difference refers to the liberation of H2 O and the addition of FA is concerted or stepwise. The calculated results have illustrated path III is more favorable. As for paths I and III, higher energy barrier of path III than path I (by 6.12 kcal mol−1 ) suggesting the formation of DMM2 is more difficult than DMM, which is in good accordance with the experimental observations [3,9]. Based on the discussion of the four pathways, we speculate the formation of DMMn from MeOH and FA will occur along the following hemiacetal-carbocation mechanism: MeOH firstly reacts with FA leading to the formation of methyl hemiformal, and then methyl hemiformal can react with MeOH leading to the formation of DMM or dehydrate to generate C1 + carbocation which can be stabilized by ILs. If introducing FA, the carbocation chains will grow and if providing MeOH the carbocation chains can be capped resulting in the formation of DMMn with different chain lengths. During these processes, SO3 H-FILs play crucial roles in every elementary step. On the one hand, via protons transfer, the cation and the anion promote the formation and the dehydration of methyl hemiformal, and the formation, the growth and the termination of the carbocations. On the other hand, the formed carbocations can be stabilized by O atom of cation. In a word, the cations and anions of SO3 H-FILs cooperatively promote the reaction of MeOH with FA resulting in the formation of DMMn with different chain length. It is worth noting that every process related to the reaction is reversible, suggesting the reaction of MeOH with FA is in equilibrium which is in good agreement with the investigation of Burger et al. [9,22,37,38] and the suggestion of Arvidson [39]. Additionally, as can be seen in Fig. 4, all the intermediates and transition states are energetically higher than the initial reactants, meaning that the
reactions should be carried out under mild heating conditions. A side product H2 O, which is unfavorable for the formation of DMMn , is generated during the reaction of MeOH with FA, indicating MeOH is probably not a good end-capping provider [24]. 3.3. The reaction of DMM with FA According to the results discussed above, the reaction of MeOH with FA follows a hemiacetal-carbocation mechanism. To clarify if this mechanism also applies to the reaction of DMM with FA mediated by SO3 H-FILs, a further study has been conducted. Fig. 5 lists the optimized species involved in the reaction. And Fig. 6 shows the corresponding energy surface. As can be seen in Fig. 5, the reaction starts with the formation of complex IM23 from DMM and SO3 H-FILs via a strong O H· · ·O hydrogen bond which lies 7.81 kcal mol−1 above the reaction entrance. Then, IM23 is converted into IM24 via TS23–24 with a barrier of 25.33 kcal mol−1 . In TS23–24 , the H atom of SO3 H group is migrating to O1 atom of DMM along the initially formed hydrogen bond making the C1 O1 bond be elongated from 1.450 Å to 2.246 Å. Simultaneously, the distance between C1 and O2 atoms are reducing from 3.116 Å to 2.206 Å. The transition vector indicated by vibration analysis as well as the subsequent IRC calculations manifest the formation of MeOH and carbocation C1 + which is stabilized by HSO4 − (see IM24). Next, FA gradually approaches C1 + leading to the chain increased carbocation C2 + which is stabilized by O atom of [BsMim]+ cation, as depicted by the process of IM25 → TS25-26 → IM26 shown in Fig. 5. The overall barrier for this step is 28.15 kcal mol−1 . The following step of the reaction refers to an end capping process of IM26 by MeOH forming DMM2 with TS27–28 as the transition state. And the barrier to be surmounted is 10.75 kcal mol−1 . However, if FA is presented, it may react with IM26 resulting in the formation of chain length increased carbocation C3 + with a barrier of 17.42 kcal mol−1 , as the process IM29 → TS29–30 → IM30 illustrated in Figs. 5 and 6. C3 + can be terminated with a lower energy barrier 6.94 kcal mol−1 if MeOH is introduced, as depicted by IM30 → TS30–31 → IM31. The chain length of C3 + will continually increase if keeping on introducing FA. Above results clearly indicate the formation of DMMn from DMM and FA is carried out via a carbocation mechanism: DMM is protonated to generate carbocation C1 + firstly, then adding FA one by one to C1 + will make the carbocation chain increase, and end-capping the formed carbocations by MeOH will result in the generation of DMMn with different chain length. For clarity, the formation process of DMMn derived from DMM and FA are sum-
234
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
Fig. 5. Optimized structures for the intermediates and transition states included in the reaction of DMM with FA catalyzed by SO3 H-FILs.
Fig. 6. Energy surface for the reaction of DMM with FA catalyzed by SO3 H-FILs.
Scheme 3. The formation process of DMMn from DMM and FA.
marized in Scheme 3. According to these results, it is probably to conclude that the formation of DMMn (n ≥ 2) is easily to achieve. However, for the same carbocation, the selectivity for FA addition and MeOH end-capping processes is different. The barriers to be surmounted for end-capping processes by MeOH to form DMM2 and DMM3 (10.75 and 6.94 kcal mol−1 for process IM26 → IM27 → TS27–28 → IM28 and IM30 → TS30–31 → IM31, respectively) are always lower than those of FA addition processes (17.42 kcal mol−1 for process IM29 → TS29–30 → IM30 and
17.06 kcal mol−1 for process IM32 → TS32–33 → IM33), suggesting the end-capping process by CH3 OH is prior to the FA addition step, which is in good agreement with the investigation of Zhao et al. [23]. Moreover, the reverse reactions for FA addition processes are more and more favorable than their forward process with the chain increase, further demonstrating the propagation of carbocation is more and more difficult. When comparing the energies of the carbocations with different lengths, such as IM26, IM30, and IM33, it is noticeably found that the longer the chain the less sta-
0.10 0.08 0.06 0.04 0.02 0.00 DMM1 7h DMM2 6h DMM4 5h DMM5 4h DMM6 3h DMM7 2h DMM8 1h DMM9 20min DMM10 DMM11 0min
Molar distribution (mol/g)
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
Fig. 7. DMM1–2,4–11 molar distributions resulting from the decomposition of DMM3 promoted by SO3 H-FILs at 30 ◦ C and 0.1 MPa. The reaction time is 7 h.
ble the carbocation. All of these factors, i.e., high selectivity for end-capping process, less favorable addition procedure of FA, and reduced stability of carbocation with long chains, contribute to the lower production of DMMn with longer chains. These results can be used to understand the experimental observations that the production of DMMn is always lower than that of DMMn−1 . In addition, we find another interesting phenomenon that the barriers about the reverse process of every elementary step are always lower or comparable with the forward processes, indicating the reverse reactions are energetically more favorable than the forward reactions. For example, the calculated results indicate the dissociation of DMM2 to regenerate DMM only requires 27.21 kcal mol−1 to surmount the highest transition state TS27–28 , which is easier by 1.91 kcal mol−1 than its formation by DMM reacting with FA. This result is in good accordance with the experimental observations that the DMM content is always higher than that of DMM2 [3,9]. Similarly, the formation of DMM3 demands more energy to surmount the barriers than its decomposition. Hence, we speculate the temperature required for DMMn (n > 1) decomposition will be lower than that for its formation. To check out if the conclusions obtained above apply to the real reaction system, we have examined the decomposition of DMM3 mediated by SO3 H-FILs at 30 ◦ C and 0.1 MPa. The product molar distribution is shown in Fig. 7. It is interesting to state that DMMn with different chain lengths were observed at the initial stage of the reaction. Moreover, the content of DMM is the highest, while DMM2,4–11 decrease gradually. However, the reaction of DMM with TOX or PF catalyzed by SO3 H-FILs hardly occurs under same conditions. It should be carried out at least 80 ◦ C and 0.1 MPa. All of these results confirm the calculated results that the decomposition of DMMn (n > 1) is easier to carry out than its formation. In addition, the calculated results that the formation of DMMn with longer chain is easy to carry out in spite of different preference have also been validated by present experimental results. Based on the calculated results, during the decomposition of DMM3 , the formation of IM26 is the most demanding step which requires 26.06 kcal mol−1 to surmount the energy barrier, as the backward process DMM3 + IL → IM26 illustrated in Fig. 6. Once this step is passed through successfully, the following steps started with IM26 which refer to the formation of DMM1,2 would proceed rapidly. Especially, the formation of DMM is energetically more favorable by 0.97 kcal mol−1 than that of DMM2 , indicating DMM is preferred to generate to DMM2 . Thus, no overshoot in the con-
235
centrations of DMM2 will appear. For the formation process of DMM4 , it is found that the generation of carbocation C4 + via TS32–33 is isoenergetic with the FA release process of DMM3 (see IM30 → TS29–30 → IM29). In combined with the more favorable backward reaction of C4 + generation process, we speculate there also no overshoot exists in the concentrations of DMM4 . All of these results are consistent with the experimental findings shown in Fig. 7. Another interesting question refers to the formation of DMMn with different lengths at the same time or in sequence? Burger et al. [9] and Zheng et al. [10] have discussed this question in their reports, but contrast conclusions were obtained. Burger et al. thought DMMn were formed simultaneously while Zheng et al. considered DMMn were synthesized in sequence. Our present experimental results which were shown in Fig. 7 indicated that DMM1–2,4–10 were simultaneously detected at the initial stage. Our calculated results demonstrate once the decomposition of DMM3 to form carbocation C2 + (DMM3 + IL → IM26) can be achieved successfully the other steps would occur rapidly in spite of a different preference (see Fig. 6). Additionally, the way for the carbocation chain propagation and the termination process of these chains indicated that the formation of DMMn was not derived from DMMn−1 except the dissociation cases of DMMn−1 resulting in the formation of DMMn . In other words, all DMMn (n ≥ 2) could be produced directly from DMM via FA insertion rather than forming DMMn−1 firstly and then yielding DMMn from DMMn−1 (see Scheme 3). Based on these results, we conjecture the formations of DMMn are probably at the same time. We hope this question be further elucidated in our future kinetic studies. During the whole process, SO3 H-FILs play important roles. Firstly, the H atom of SO3 H-FILs is “borrowed” by DMM to release a MeOH and a carbocation, and then the formed carbocation can be stabilized by O atom of SO3 H-FILs. At last, the H atom is “returned” to SO3 group during the termination process making the regeneration of SO3 H-FILs. These results indicate both cations and anions affect the catalytic activity of ionic liquids on the formation of DMMn . Additionally, it has been confirmed that many protic ionic liquids possess brönsted acidic nature, such as [HMIM]+ [HSO4 ]− [40]. So it will be valuable to investigate the effects aroused by different cations, anions and protic ionic liquids on the formation of DMMn . However, this is another issue which beyond the scope of present work and will be pursued in our future work. 4. Conclusions The detailed formation mechanism of DMMn from MeOH (DMM) with TOX or with PF promoted by SO3 H-FILs has been examined using DFT calculations. The following conclusions can be obtained: (1) Before reacting with MeOH or DMM, TOX and PF will decompose into monomeric FA firstly. The decomposition of TOX follows a two-step mechanism: TOX protonated by SO3 H-FILs firstly leading to the formation of the ring-opened trioxymethylene chain, and then the transferred hydrogen atom goes back to SO3 H-FILs realizing the formation of three FA monomers. As for PF, its decomposition to produce monomeric FA is carried out by one-step mechanism. (2) When MeOH is chosen as the capping provider, the condensation process follows a hemiacetal-carbocation mechanism: MeOH reacting with FA results in the formation of hemiacetal which is a key intermediate for the reaction. This hemiacetal can react with MeOH leading to the generation of DMM or react with FA to bring out the production of chain increased carbocations. Capping these carbocations by MeOH will obtain
236
F. Wang et al. / Journal of Molecular Catalysis A: Chemical 408 (2015) 228–236
the DMMn . While DMM is selected to terminate the end, a carbocation mechanism will be preferred. With the assistant of SO3 H-FILs, DMM firstly dissociates into MeOH and a carbocation. Inserting FA into this carbocation one by one will generate the chain grown carbocations which can be terminated by MeOH leading to the formation of DMMn . (3) The product distribution of DMMn that DMMn < DMMn−1 can be understood by the following points: 1) for the same carbocation, its termination by MeOH is always more favorable than its chain growth process by the addition of FA; 2) the reverse reaction of FA addition step proceeds more rapidly than the forward reaction which brings difficulty for the carbocation growth; 3) the stability of carbocations reduces with the increase of chain lengths; 4) the formation of DMMn is more difficult than its decomposition. (4) SO3 H-FILs play vital roles no matter in the decomposition of TOX/PF or in the reaction of FA with MeOH/DMM. On the one hand, the cations and anions of SO3 H-FILs can synergistically promote the reaction via protons transfer. On the other hand, the formed carbocation can be stabilized by SO3 H-FILs via forming C O bonds. In addition, the intermediates and transition states can also be stabilized through their hydrogen-bonding interactions with SO3 H-FILs. (5) Both the decomposition of TOX/PF and the reaction of FA with MeOH/DMM are equilibrium process. Moreover, the reactions should be operated under heating conditions.
[3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20] [21] [22] [23] [24] [25]
Acknowledgements This work was supported by the National Basic Research Program of China (973 Program, No. 2011CB201404), and National Natural Science Foundation of China (Project No. 21203218, 21133011, 21473225). We gratefully acknowledge the computing resources and time on the Supercomputing Center of Cold and Arid Region Environment and Engineering Research Institute of Chinese Academy of Sciences. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.molcata.2015.07. 028 References [1] F. Willian, E. Richard, US: 2449469 (1948). [2] F. Zhao, H.J. Li, H.L. Song, C.G. Xia, Nat. Gas Chem. Ind. 1 (2013) 1–6.
[26] [27] [28] [29] [30] [31] [32] [33] [34] [35] [36] [37] [38] [39] [40]
H.J. Li, H.L. Song, L.W. Chen, C.G. Xia, Appl. Catal. B 165 (2015) 466–476. P.H. Gary, J.S., Michael, US: 6160174A (2000). P.H. Gary, J.S., Michael, US: 2003/0171534A1 (2003). P.H. Gary, J.S., Michael, US: 6437195 B2 (2002). L. Wang, W.T. Wu, T. Chen, Q. Chen, M.Y. He, Chem. Eng. Commun. 201 (2014) 709–717. T. Chen, L. Wang, Q. Chen, M.Y. He, Ion Exc. Ads. 28 (2012) 456–462. ¨ J. Burger, E. Strofer, H. Hasse, Ind. Eng. Chem. Res. 51 (2012) 12751–12761. Y.Y. Zheng, Q. Tang, T.F. Wang, Y.H. Liao, J.F. Wang, Chem. Eng. Technol. 36 (2013) 1951–1956. Q. Zhao, H. Wang, Z.F. Qin, Z.W. Wu, J.B. Wu, W.B. Fan, J.G. Wang, J. Fuel Chem. Technol. 39 (2011) 918–923. X.H. Zhang, P.T. Wu, Y. Zhang, Z.C. Jiang, T.T. Liu, X.Y. Gao, Y.Q. Hu, Chem. React. Eng. Technol. 30 (2014) 140–144. J.Q. Zhang, D.Y. Fang, D.H. Liu, Ind. Eng. Chem. Res. 53 (2014) 13589–13597. J. Chen, Z.H. Tang, C.G. Xia, X.Z. Zhang, Z. Li, CN: 101182367A (2008). J. Chen, H.Y. Song, C.G. Xia, X.Z. Zhang, Z.H. Tang, CN: 101665414A (2010). J. Zhang, M. Shi, D. Fang, D. Liu, React. Kinet. Mech. Catal. 113 (2014) 459–470. H.J. Li, H.L. Song, F. Zhao, L.W. Chen, C.G. Xia, J. Energy Chem. 24 (2015) 239–244. J. Chen, H.Y. Song, C.G. Xia, Z. Li, CN: 102249869A (2011). R. Patrini, M. Marchionna, EP: 1505049A1 (2005). R. Patrini, M. Marchionna, US: 6534685 (2003). H. Schelling, E. Ströfer, R. Pinkos, A. Haunert, G. Tebben, H. Hasse, S. Blagov, WO: 2006045506 A1, (2006). E. Ströfer, H. Hasse, S. Blagov, US: 20080207954 A1 (2008). Y.P. Zhao, Z. Xu, H. Chen, Y.C. Fu, J.Y. Shen, J. Energy Chem. 22 (2013) 833–836. J. Burger, M. Siegert, E. Ströfer, H. Hasse, Fuel 89 (2010) 3315–3319. M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery, J.E. Peralta Jr., F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, O. Farkas, J.B. Foresman, J.V. Ortiz, J. Cioslowski, D.J. Fox, Gaussian 09, Revision B.01, Gaussian, Inc., Wallingford, CT, 2010. S. Grimme, J. Comp. Chem. 27 (2006) 1787. W.J. Hehre, R. Ditchfield, J.A. Pople, J. Chem. Phys 56 (1972) 2257–2261. P.C. Hariharan, J.A. Pople, Theor. Chim. Acta 28 (1973) 213–222. G.W. Spitznagel, T. Clark, J. Chandrasekhar, P.v.R. Schleyer, J. Comput. Chem. 3 (1982) 363–371. T. Clark, J. Chandrasekhar, G.W. Spitznagel, P.v.R. Schleyer, J. Comput. Chem. 4 (1983) 294–301. C. Peng, H.B. Schlegel, Isr. J. Chem. 33 (1993) 449–454. H.B. Schlegel, J. Comput. Chem. 3 (1970) 214. K. Fukui, J. Phys. Chem. 74 (1970) 4161–4163. H.Y. Song, Z. Li, J. Chen, C.G. Xia, Catal. Lett. 142 (2012) 81–86. A.C. Cole, J.L. Jensen, I. Ntai, K.L.T. Tran, K.J. Weaver, D.C. Forbes, J.H. Davis, J. Am. Chem. Soc. 124 (2002) 5962. J.M. Tatibouët, L.P. Hélène, J. Mol. Catal. A 171 (2001) 205–216. J. Burger, H. Hasse, Chem. Eng. Sci. 99 (2013) 118–126. J. Burger, E. Ströfer, H. Hasse, Chem. Eng. Res. Des. 91 (2013) 2648–2662. M. Arvidson, M.E. Fakley, M.S. Spencer, J. Mol. Catal. 41 (1987) 391–393. T.L. Greaves, C.J. Drummond, Chem. Rev. 108 (2008) 206–237.